Dear Travis,
Actually, saving the current cluster works fine so you can just choose
the current location to be within your cluster of interest followed by:
* Save > current cluster: cluster_mask.gii
* m = gifti('cluster_mask.gii');
* g = gifti('beta_0001.gii');
* mean(g.cdata(find(m.cdata)))
Best regards,
Guillaume.
On 12/04/17 10:38, Guillaume Flandin wrote:
> Dear Travis,
>
> There might be a direct functionality for this in CAT12 but you can do
> it manually in SPM12:
>
> * Save > all clusters (n-ary): clusters_mask.gii
> * m = gifti('clusters_mask.gii');
> % compute cluster sizes
> histc(m.cdata,0:max(m.cdata))
> 20147 113 47 43 ...
> % First value is background
> * If I'm interested in the mean beta_1 value within the cluster with 113
> vertices, I would do:
> g = gifti('beta_0001.gii');
> mean(g.cdata(find(m.cdata==1)))
>
> spm_summarise should do this for you automatically but this function
> hasn't been updated to handle surface data yet, and it seems that the
> option to save a mask of the current cluster does not work as expected
> hence relying on n-ary here.
>
> Best regards,
> Guillaume.
>
>
> On 11/04/17 21:39, Travis Beckwith wrote:
>> Hi,
>>
>> I've been using the CAT12 toolbox to generate cortical thickness estimates for analyses. Does anyone know of a way to use a custom mask to extract values from the analysis output files? Specifically, I would like to make a mask from a statistically significant cluster and then use that mask as an ROI for extracting the beta values from the beta_000*.gii files to examine how much confounding variables change the beta values. I understand how to do this for a VBM analysis in SPM when the output are nifti files, but I'm having trouble with the gifti files.
>>
>> Thanks,
>>
>> Travis
>>
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|